基于边缘的多尺度立体匹配算法  

Segment-based Stereo Matching Using Edge Dynamic Programming

作　　者：夏林元[1,2] 肖俊[1] 林丽群[3] 

机构地区：[1]武汉大学测绘遥感信息工程国家重点实验室,武汉市珞喻路129号430079 [2]中山大学地理科学与规划学院,广州新港西路135号510275 [3]湖北大学资源与环境学院,武汉市学院路11号430062

出　　处：《武汉大学学报（信息科学版）》2011年第7期767-770,共4页Geomatics and Information Science of Wuhan University

基　　金：国家973计划基金资助项目(2006CB701301)

摘　　要：提出了一种面向对象的多尺度递进的立体匹配算法,实现了由粗到细的快速立体匹配。首先对参考图进行粗分割获取目标对象,针对目标区域采用边缘线动态规划的方法获取视差,根据精度需求进一步判断该目标是否需要细分以获取更加精细的视差,得到有效的视差图。实验结果表明此方法快速有效。Most current stereo matching method is based on pixel processing,but object-oriented approach is more conform to the human visual habit.This paper put forward a scale-span multi-scale stereo matching method,achieved by coarse to fine stereo matching.Firstly the reference image is rough segmented by an object-oriented segmentation method,then by employing a refined DP algorithm which is based on the inner edgeline of the segment,disparity are obtained.According to the accuracy demand,determine whether need to do the refined segmentation.Finally,the final disparity images are obtained.The experimental results indicate that the algorithm has high matching precision and fast calculating speed.

关 键 词：立体匹配 图像分割 动态规划 多尺度 

分 类 号：P237.3[天文地球—摄影测量与遥感]